UX Camp Winter 2024
ChatGPT: UX Research Friend, Foe, or Both?
ChatGPT and other large language models (LLMs) seem like magical question answerers. So magical that they make some people wonder: why should I ask real people about my product, when I can ask ChatGPT what they think instead? Why do I need a researcher if an LLM seems to do research for me? And, if ChatGPT can do research, do we still need human UX researchers? LLMs’ vortex of questions are coming to our profession, and making some of us (me included) a tad anxious.
In this talk, I’ll show how UX researchers can see ChatGPT not as a research participant—or competitor—but as a collaborator. In particular, I’ll share:
- How LLMs actually work, and how they’re more like “persona builders” than human researchers
- The real strengths of LLMs that can help you plan more creative interviews and usability tests
- The human strengths you can bring to research analysis/synthesis that LLMs can’t replicate (and probably never will)
- Two ways LLMs can inadvertently lead humans astray – and how to avoid them when you use something like ChatGPT in research
You’ll come away with concrete tactics for appropriately incorporating LLMs into your UX research process, as well as ways to explain their limitations to over-enthusiasts. And hopefully some of you, like me, will learn to worry about LLMs a little less.
